Regression-Based Strategy for Defining Subgroups in a Social Experiment: MDRC Working Papers on Research Methodology

摘要

The evidence and discussion in this paper strongly support the idea that accounting for the heterogeneity of students in the Career Academies Evaluation is an important element of any strategy designed to assess the impact of the Academies on the diverse group of students they serve. Impact estimates which aggregate results across students with different academic trajectories conceal a substantial amount of variation across students in the effects of the Academies on key outcomes. Therefore, in order to assess the effects of Career Academies more sensitively, it is necessary to develop a strategy for differentiating among students who, in the absence of the Academy treatment, would experience different academic outcomes.